CN114514082A

CN114514082A - Techniques for analyzing sensor data in powder bed additive manufacturing

Info

Publication number: CN114514082A
Application number: CN202080067499.1A
Authority: CN
Inventors: 迪特尔·施瓦策; 金姆·库林; 安德里亚斯·霍佩; 贝贝尔·克雷茨; 丹尼尔·艾伯茨
Original assignee: SLM Solutions Group AG
Current assignee: Nikon Slm Solutions Co ltd
Priority date: 2019-09-25
Filing date: 2020-08-25
Publication date: 2022-05-17
Anticipated expiration: 2040-08-25
Also published as: US11685122B2; JP2022544629A; CN114514082B; WO2021058216A1; JP7202510B2; EP4034324A1; US20220194011A1

Abstract

An apparatus for analyzing sensor data of a sensor provided in an apparatus for manufacturing a three-dimensional workpiece by irradiating a layer of a raw material with an energy beam is provided. The device includes a control unit configured to receive sensor data as a time series of data values. Each data value represents a process condition within the apparatus during fabrication of the three-dimensional workpiece. The control unit is further configured to receive planning data for the three-dimensional workpiece. The planning data defines a plurality of scan vectors and an order in which the energy beam is scanned along the scan vectors. The control unit is further configured to associate the time series of data values with respective vector data of the scan vectors of the planning data to form a plurality of sets of data values for the respective scan vectors, and to define a set of at least two of the plurality of scan vectors based on the planning data. The scan vectors of the set satisfy a predetermined similarity criterion. The control unit is further configured to compare the set of data values of the first scanning vector in the set with the set of data values of the at least one second scanning vector in the set, or the set of data values of the first scanning vector in the set with a combined set of data values derived from at least two second scanning vectors of the set, and to determine a quality metric of the workpiece at the position of the first scanning vector based on the comparison. Corresponding methods and computer program products are also provided.

Description

Techniques for analyzing sensor data in powder bed additive manufacturing

Technical Field

The present invention relates to a technique for analyzing sensor data of a sensor provided in an apparatus for manufacturing a three-dimensional workpiece by irradiating a layer of a raw material with an energy beam. The techniques may be embodied in at least one of an apparatus, a method, and a computer program product. The apparatus for producing a three-dimensional workpiece may be an apparatus for selective laser melting or selective laser sintering.

Background

Powder bed melting is an additive layering process by which powdered, especially metallic and/or ceramic, raw materials can be processed into three-dimensional workpieces of complex shape. For this purpose, a raw material powder layer is applied to a carrier and is irradiated (e.g. laser or particle irradiation) in a site-selective manner according to the desired geometry of the workpiece to be produced. The radiation penetrates the powder layer so that the raw material powder particles are heated and thus melted or sintered. The layers of raw material powder are then applied in succession to the layers on the carrier which have been subjected to the radiation treatment until the workpiece has the desired shape and size. Powder bed fusion can be used, for example, to manufacture prototypes, tools, replacement parts, high value components, or medical prostheses, such as dental or orthopedic prostheses, based on CAD data. Examples of powder bed melting techniques include selective laser melting and selective laser sintering.

Apparatus for manufacturing one or more workpieces according to the above-described techniques are well known. For example, EP2961549a1 and EP2878402a1 describe devices for producing three-dimensional workpieces according to the selective laser melting technique. The general principles described in these documents are also applicable to the techniques of this disclosure. However, the present disclosure is not limited to selective laser melting techniques and may be applied to other additive manufacturing processes according to which a layer of raw material may be solidified by a beam of radiation.

Furthermore, it is also known to monitor attributes of a manufacturing process by providing at least one sensor that continuously measures a particular attribute (e.g., a physical condition) during the manufacturing process. For example, a plurality of sensors may be provided to measure the temperature within the manufacturing chamber of the apparatus, the oxygen content within the manufacturing chamber, the inert gas pressure, and/or the melt pool temperature, i.e., the melt pool temperature caused by the radiation beam impinging on the raw material. In particular, it is known to provide a so-called melt pool monitoring system which includes a sensor in the form of a pyrometer which measures the intensity of the thermal radiation generated in the melt pool. More specifically, a pyrometer receives thermal radiation emitted from the melt pool and the pyrometer is configured to measure at least one intensity value of the thermal radiation. The pyrometer (or a control unit connected to the pyrometer) may determine the temperature of the molten bath based on the measured intensity values.

The data values resulting from the measurements of the respective sensors may be stored in a memory of a control unit of the apparatus, it being known, for example, to represent layers of the manufactured workpiece in a two-dimensional image of the generated data values. On the basis of the two-dimensional image, it is possible, for example, to identify regions of the layer of the workpiece in which the sensor data differs significantly from other regions. This may indicate that in the identified region, the quality of the workpiece is different from the quality of the other regions.

However, by performing this form of data evaluation, valuable information is lost and cannot be considered in the data evaluation. In particular, individual data points cannot be assigned to a particular scan vector of the generated workpiece. Another drawback of the above known form of data evaluation is that when evaluating the data irradiation of the entire layer has already been performed, it is not possible to monitor the irradiation on-line during irradiation of the layer, nor to adjust the irradiation parameters or stop the irradiation process based on the evaluation during irradiation of the layer.

Disclosure of Invention

Accordingly, the present invention is directed to an apparatus and method that solves the above-mentioned problems and/or other related problems. In particular, it is an object of the present invention to provide an improved technique for analyzing sensor data of a sensor.

This object is achieved by an apparatus according to claim 1, an apparatus according to claim 13, a method according to claim 14 and a computer program product according to claim 15.

According to a first aspect, an apparatus for analyzing sensor data of a sensor provided in a device for manufacturing a three-dimensional workpiece by irradiating a layer of a raw material with an energy beam is provided. The device includes a control unit configured to receive sensor data as a time series of data values. Each data value represents a process condition within the apparatus during fabrication of a three-dimensional workpiece. The control unit is further configured to receive planning data for the three-dimensional workpiece. The planning data defines a plurality of scan vectors and an order in which the energy beam is scanned along the scan vectors. The control unit is further configured to associate the time series of data values with respective vector data of the scan vectors of the planning data to form a plurality of sets of data values of the respective scan vectors, and to define a set of at least two of the plurality of scan vectors based on the planning data. The scan vectors in the set of scan vectors satisfy a predetermined similarity criterion. The control unit is further configured to compare the set of data values of the first one of the set of scan vectors with the set of data values of the at least one second one of the set of scan vectors, or the set of data values of the first one of the set of scan vectors with a combined set of data values derived from at least two second one of the set of scan vectors, and to determine a quality metric of the workpiece at the position of the first scan vector based on the comparison.

The apparatus may be, for example, a standalone apparatus configured to receive sensor data from an apparatus for manufacturing three-dimensional workpieces, or the apparatus may be part of a control unit of an apparatus for manufacturing three-dimensional workpieces. The time series of data values may be provided, for example, as a data stream or in the form of a data file. One or more of the data values may be assigned a timestamp such that each of the data values is assigned a particular point in time (e.g., when the sampling frequency of the sensor is known). For example, the timestamp may indicate a point in time at which the manufacturing process of the workpiece has begun. The data values may be the result of sampling the corresponding sensor at a constant sampling rate, such that the spacing between the individual data values is known and constant. The process conditions may be physical process conditions and may be measured, for example, within a process chamber of an apparatus for fabricating a three-dimensional workpiece. Examples of process conditions that may be measured include, for example, the temperature within the manufacturing chamber, the oxygen content in the manufacturing chamber, the inert gas pressure, and/or the melt pool temperature.

The planning data may be provided in the form of a data file. The planning data may be the result of a process according to which a CAD file of the workpiece is analyzed, and wherein the workpiece to be generated is subdivided into a plurality of layers, and each of the plurality of layers is subdivided into a plurality of scan vectors. In general, the scan vector may be defined as the path scanned by the radiation beam of the apparatus between the point at which the radiation beam starts to generate the melt pool and the point at which the radiation beam stops generating the melt pool. Thus, according to this definition of the scan vector, the scan vector may be curved and the direction of the scan vector may be changed. The scan vectors are defined as the portions of the path scanned by the radiation beam that form a straight line, in accordance with the more rigorous definition of the scan vectors, as is applicable to the teachings disclosed herein. In this case, for example, a layer of the workpiece to be generated may be irradiated using a fill pattern comprising a plurality of parallel scan vectors. Each scan vector has a certain predetermined length, and a plurality of scan vectors having the same length may be provided, but a plurality of scan vectors having different lengths may also be provided. For example, a checkerboard pattern may be filled in an interior region of a workpiece, where each region of the pattern is filled with a plurality of parallel scan vectors, and the plurality of scan vectors of different regions point in different directions.

As described above, the planning data defines a plurality of scan vectors and an order in which the energy beam is scanned along the plurality of scan vectors. The planning data may also define other information and/or data related to the manufacturing process. For example, the planning data may include one or more of the following: a number of scan vectors planned per layer, a positioning of one or more workpieces, a scan speed (e.g., a scan speed of a respective scan vector), a laser power (e.g., for a respective scan vector), a focus size (e.g., for a respective scan vector), a scan direction of a respective scan vector, and a scan sequence of a respective scan vector.

Each of the plurality of sets of data values corresponds to a particular scan vector. For example, each data set may be provided in the form of an n-dimensional vector, where n is the number of sensor values measured during irradiation of the corresponding scan vector. When the sampling rate of the sensor is known, the time to scan the scan vector can be calculated. Further, when the scan speed is known (e.g., from planning data), a respective position related to a layer of the workpiece may be calculated for each sensor value (i.e., for each data point of the data set).

A group comprising at least two scan vectors is defined based on the planning data such that the scan vectors of the group of scan vectors satisfy a particular similarity criterion. In other words, the scan vectors of the set are similar for at least one criterion. More precisely, for each of a plurality of scan vectors for planning data, a respective similarity value may be calculated with respect to each of the remaining scan vectors. For example, if the number of scan vectors is n, n-1 similarity values may be calculated for each of the n scan vectors. Scan vector pairs with similarity values below a predetermined threshold may be grouped into the same group.

In addition to the group comprising at least two scan vectors, one or more other groups may be defined. The vectors in the one or more further groups each satisfy a different further predetermined similarity criterion compared to the vectors of the (first) group.

In the comparison, the set of data values of the first scan vector of the group may be compared with the set of data values of the one or more second scan vectors of the same group. Alternatively, the set of data values of the first scan vector is compared with a combined set of data values deduced from at least two second scan vectors of the same group. The first scan vector in the set may be the "youngest" scan vector in the set, i.e., the most recently irradiated scan vector. The set of data values of the first scan vector may be compared with only one other scan vector (e.g., a previously irradiated scan vector), and the result of this comparison may be compared with the previous comparison result of the set of data of scan vectors within the same set (i.e., within the set of first scan vectors). If the result of the comparison of the set of data values of the first scan vector with the set of data values of the further scan vector differs significantly (i.e. above a predetermined threshold) from the previous result of the comparison, it may be determined that the quality of the workpiece at the position of the first scan vector has changed. In other words, the likelihood of a change in mass relative to other parts of the workpiece at the corresponding location of the workpiece is increased.

Similarly, the set of data values of the first scanning vector is compared not only with the set of data values of one second scanning vector, but also with the set of data values of a plurality of second scanning vectors, e.g. with each of the remaining (second) scanning vectors of the set. Alternatively, a combined set of data values may be deduced from the at least two second (remaining) scan vectors, wherein a comparison of the set of data values of the first scan vector is compared with the deduced combined set of data values. For example, the combined set of data values may be an averaged set of data values. More precisely, an averaging form may be applied to the set of data values of the at least two second scanning vectors of the set to derive an averaged set of data values. Averaging may comprise deriving an arithmetic mean of the data values of the different vectors. The combined set of data values may be deduced by considering at least one of meridians (meridians), extrema and quantiles.

Based on the comparison, a quality metric of the workpiece at the location of the first scan vector is determined. For example, in the case where the comparison results in a height difference, it may be determined that the mass of the workpiece at the location of the first scan vector is lower than the mass of the remainder of the workpiece.

By implementing the techniques described above, sensor data for multiple similar scan vectors may be compared to one another. Thus, an "abnormal" or "problematic" data value may be identified by comparing the data value to data values of similar scan vectors, where it may be assumed that the data values of scan vectors in the same group do not differ too much from each other. In some cases, comparison with a reference value or a reference set of data values may be avoided. Rather, there is no need to compare with so-called "actual truth". However, in addition to the group-wise comparison disclosed herein, a comparison to a reference value may also be performed in order to obtain additional information and/or verification.

The data value may be an intensity value or a temperature value of thermal radiation generated in a melting bath in which the energy beam is irradiated onto the raw material.

In other words, the data value may be an intensity value of thermal radiation generated in the melting bath in which the energy beam irradiates the raw material, or the data value may be a temperature value of thermal radiation generated in the melting bath in which the energy beam irradiates the raw material. In both cases, the sensor may be a pyrometer. The pyrometer may be configured to detect thermal radiation intensity in one or more wavelength bands. The one or more thermal radiation intensities may then be output as intensity values, or corresponding temperature values may be calculated based on the one or more thermal radiation intensities. Thus, the data values may be intensity values (directly), or the data values may be temperature values derived from one or more thermal radiation intensity values. Since the intensity of the thermal radiation emitted from the melt pool can be varied rapidly, the sampling frequency of the sensor can be high enough to record multiple data values over the course of one scan vector.

The apparatus may be configured to perform at least the steps of receiving sensor data, correlating, comparing, and determining during a manufacturing process of the workpiece.

For example, sensor data may be measured continuously during the manufacturing process, thereby generating new data values during the process. To process these newly generated sensor data, the foregoing steps may be performed during the manufacturing of the workpiece. Thus, the sensor data may be analyzed in near real-time, i.e., as the sensor data is generated. Thus, the deviation can be quickly reacted to by adjusting parameters of the manufacturing process, for example, according to a closed-loop control.

The control unit may also be configured to adjust a process parameter of the device during the manufacturing process based on the determined quality metric.

The process parameters may be, for example, laser power, laser focus position, scanning speed, etc. In this way, countermeasures that help cope with the degradation of the quality of the workpiece can be initiated.

The control unit may be further configured to associate the time data with a time series of data values.

The time data may comprise at least one time stamp associated with a respective data value of the time series. In case the sampling frequency of the sensor is known, it may be sufficient to associate only one time data point with the corresponding data value of the sensor. Thus, each of the data values may be assigned a respective point in time. The time data may indicate the time of day and optionally the date. Additionally or alternatively, the time data may indicate a time relative to a predefined reference time t-0 (e.g., the start of a manufacturing process). For each data value, a corresponding location within the workpiece may be identified based on the time data.

The similarity criterion may consider at least one of the following: length similarity of scan vectors, direction similarity of scan vectors, and whether a scan vector is part of a workpiece profile.

In other words, a plurality of scan vectors having similar lengths may be grouped together. Thus, respective groups may be provided for a plurality of intervals of different lengths. Similarly, multiple scan vectors with similar directions may be grouped together. A plurality of scan vectors that are part of the workpiece profile may be grouped together.

The control unit may be further configured to compare the set of data values of the first one of the set of scan vectors with a predefined set of reference data values associated with the set.

The reference set of data values may be stored in a memory of the control unit or may be calculated in real time. The reference set may be pre-calculated or predefined with respect to the attributes of the corresponding group. For example, where it is known that a scan vector having a particular length (or within a particular range of lengths) should exhibit a particular typical response to measured sensor values, the reference set may include these typical values. Furthermore, the reference set may be deduced from at least one previous measurement (i.e. a measurement during an earlier manufacturing process performed using the same device). In this case, for example, the reference set may be a set of average data values of corresponding groups of scan vectors that satisfy the same similarity criterion.

The control unit may be configured to compare the set of data values of the first scanning vector in the group with the set of data values of the second scanning vector previously irradiated in the group and to determine at least one difference value representing a difference between the sets of data values of the first scanning vector and the second scanning vector based on the comparison.

The difference value may be a distance measure of the set of compared data values. In the manner described above, the data values of these vectors can be compared one by one. In other words, each time a new scan vector is irradiated and corresponding data values are recorded, these data values are compared with the data values of the previous scan vectors in the same group. The result of the comparison may then be compared, for example, to a threshold value. In the case where the difference exceeds a predetermined threshold, it can be determined that the new scan vector (first scan vector) includes an abnormal data value that may cause a deviation in the quality of the workpiece.

The control unit may be further configured to compare at least one difference value with a stored difference value representing a difference between sets of data values of two previously irradiated scan vectors.

Thus, a typical degree of difference between the respective vectors may be taken into account, and it may be determined that the first scanning vector comprises an abnormal data value only if the difference value of the first scanning vector exceeds the typical degree.

The control unit may be configured to perform the comparing step by taking into account at least one of: a series of absolute data values (e.g. a decrease and/or an increase of data values from one scan vector to another scan vector), a relative comparison (e.g. a difference, a deviation and/or a fluctuation from each other) of the data values, a correlation of the data value set of the first scan vector with the data value set of the at least one second scan vector or a correlation of the data value set of the first scan vector with a combined data value set deduced from the at least two second scan vectors, an absolute difference of the data values of the first scan vector with the data values of the at least one second scan vector or an absolute difference between the combined data value sets deduced from the at least two second scan vectors for the data values of the first scan vector, and an extreme value of the set of data values of the first scan vector with the set of data values of the at least one second scan vector or an extreme value of the set of data values of the first scan vector with a combined number deduced from the at least two second scan vectors And analyzing the extreme value of the set of values.

The control unit may be configured to perform the comparing step by considering only a subset of the set of data values of the first scanning vector and/or a subset of the set of data values of the at least one second scanning vector, or considering a subset of the combined set of data values derived from the at least two second scanning vectors.

For example, a subset of the set of data values of the first scanning vector may start at a predetermined time t0 after the data values of the respective vector are recorded. For example, considering only a subset of the set of data values may have the following advantages: a plurality of subsets having the same length may be compared with each other. For example, when the set of scan vectors includes vectors having different lengths and thus different numbers of data values, each vector may be clipped to a predetermined uniform length so that comparisons between vectors may be easily made. Furthermore, by considering only a subset, portions of the respective vectors that clearly include deviating data values (e.g., turning points of the vectors, vectors that end at raw material edges, etc.) may be ignored.

The control unit may be configured to output a data set representing at least one two-dimensional image of a layer of the workpiece, wherein a time-series of a plurality of data values is assigned to a corresponding plurality of pixels of the two-dimensional image.

In this way, the measurements of the sensors can be analyzed individually.

According to a second aspect, an apparatus for manufacturing a three-dimensional workpiece by irradiating a layer of a starting material with an energy beam is provided. The apparatus comprises the device of the first aspect, an energy beam source for generating and irradiating an energy beam onto a plurality of layers of raw material, and a sensor configured to measure and transmit a time series of data values to the device.

The energy beam may be, for example, a laser or an electron beam. The raw material layer may be a raw material powder layer (e.g., a metal powder or a ceramic powder). Since the apparatus of the first aspect is part of the apparatus of the second aspect, the details of the first aspect described above may also be applied to the second aspect.

According to a third aspect, a method for analyzing sensor data of a sensor provided in an apparatus for manufacturing a three-dimensional workpiece by irradiating a raw material layer with an energy beam is provided. The method includes receiving sensor data as a time series of data values. Each data value represents a process condition within the apparatus during fabrication of a three-dimensional workpiece. The method also includes receiving planning data for the three-dimensional workpiece. The planning data defines a plurality of scan vectors and an order in which the energy beam is scanned along the scan vectors. The method also includes associating the time series of data values with respective vector data of the scan vectors of the planning data to form a plurality of sets of data values for the respective scan vectors, and defining a set of at least two of the plurality of scan vectors based on the planning data. The scan vectors of the set of scan vectors satisfy a predetermined similarity criterion. The method further comprises comparing the set of data values of the first one of the set of scan vectors with the set of data values of at least one second one of the set of scan vectors, or comparing the set of data values of the first one of the set of scan vectors with a combined set of data values derived from at least two second one of the set of scan vectors, and determining a quality metric of the workpiece at the position of the first scan vector based on the comparison.

The method according to the third aspect may be performed using the apparatus according to the first aspect. The details and further explanations discussed above in relation to the first aspect may also be applied to the method of the third aspect.

According to a fourth aspect, a computer program product stored on a computer readable storage medium is provided. The computer program product comprising computer readable instructions for causing a computer to perform the method according to the third aspect.

The details of the discussion and further explanations above regarding the apparatus of the first aspect may also be applied to the computer program product of the fourth aspect.

Drawings

Preferred embodiments of the present invention are described in more detail with reference to the accompanying schematic drawings, in which:

fig. 1 shows a schematic view of an apparatus for producing a three-dimensional workpiece by irradiating a layer of a starting material with an energy beam according to the present disclosure, the apparatus comprising a sensor and a device for analyzing sensor data of the sensor;

FIG. 2 illustrates an apparatus for analyzing sensor data of a sensor disposed in a device for fabricating a three-dimensional workpiece by irradiating a layer of a raw material with an energy beam according to the present disclosure;

FIG. 3 shows a top view of an exemplary structure having multiple scan vectors compared to each other;

FIG. 4 shows a graph of two scan vectors compared to each other; and

fig. 5 shows a schematic representation of a plurality of layers, wherein the scan vectors of different layers are compared with each other.

Detailed Description

Fig. 1 shows a schematic view of an apparatus 2 for producing a three-dimensional workpiece 4. The apparatus 2 comprises a sensor 6 in the form of a sensor measuring thermal radiation and a device analyzing sensor data generated by the sensor 6. The device comprises a control unit 8, which in the embodiment of fig. 1 is a general control unit 8 of the device 2.

The apparatus 2 shown in fig. 1 is an apparatus for producing a three-dimensional workpiece 4 by selective laser melting. The apparatus 2 includes a process chamber 10. The process chamber 10 may be sealed to the surrounding atmosphere, i.e., the process chamber 10 is sealed from the surrounding environment. A powder application device 12 arranged in the process chamber 10 is used for applying the raw material powder onto the carrier 14. The vertical moving unit 16 is provided so that the carrier 14 can move in the vertical direction, so that the carrier 14 can move downward in the vertical direction when the workpieces 4 are manufactured in layers from the raw material powder on the carrier 14 as the manufacturing height of the workpieces 4 increases.

Since the movement of the carrier 14 by means of the vertical movement unit 16 is well known in the field of selective laser melting, it will not be explained in detail here. As an alternative to the movable carrier 14, the carrier 14 may be provided as a stationary (or fixed) carrier (in particular in the vertical Z-direction), wherein the irradiation unit 18 (see below) and the process chamber 10 are configured to move upwards during the manufacturing process (i.e. with an increasing manufacturing height of the work piece 4).

The support surface of the support 14 is defined as a horizontal plane (x-y plane), wherein a direction perpendicular to the plane is defined as a vertical direction (z direction). Thus, each uppermost layer of raw material powder extends in a plane parallel to the above-defined horizontal plane (x-y plane).

The apparatus 2 further comprises an irradiation unit 18, which irradiation unit 18 is used to selectively irradiate laser radiation onto the uppermost raw material powder applied on the carrier 14. By means of the irradiation unit 18, the raw material powder applied to the carrier 14 can be subjected to laser radiation in a site-selective manner, depending on the desired geometry of the workpiece 4 to be produced.

The irradiation unit 18 comprises at least one laser beam source 20 generating a laser beam 22. In an alternative embodiment, a particle beam (e.g., an electron beam) may be used to melt the raw material powder instead of the laser beam 22. The laser beam source 20 may comprise, for example, a diode-pumped ytterbium-doped fiber laser that emits laser light having a wavelength of about 1070nm to 1080 nm. The irradiation unit 18 comprises a scanning unit 24, which scanning unit 24 is used to direct a laser beam 22 onto the uppermost raw material powder applied on the carrier 14 in order to locally heat and melt the powder at a desired location, i.e. in a site-selective manner. The position of the laser beam 22 can be moved in the x-y plane of the uppermost raw material powder layer by means of the scanning unit 24. The scanning unit 24 may include one or more movable mirrors and may be configured in the form of a galvanometer scanner.

Furthermore, irradiation unit 18 may comprise other optical components in addition to scanning unit 24, such as a beam expander for expanding laser beam 22, focusing optics for focusing laser beam 22 in a direction along the beam path, and/or an objective lens. The objective lens may be an f-theta objective lens arranged in the beam path after the scanning unit 24. The operation of the irradiation unit 18 is controlled by the control unit 8. Further, the control unit 8 is configured to control other components of the apparatus 2, such as the vertical moving unit 16 and the powder applying apparatus 12.

Furthermore, the sensor 6 is arranged in the irradiation unit 18. The sensor 6 of the present embodiment is a pyrometer configured to detect thermal radiation emitted by the melt pool produced by the laser beam 22. In the melting tank, a laser beam 22 is irradiated onto the raw material powder and causes the powder to melt. A semi-transparent mirror 28 is provided which directs thermal radiation to the sensor 6. The half mirror 28 may be wavelength dependent, preferably reflecting thermal radiation (light in the thermal radiation wavelength region) to the sensor 6. More precisely, the semi-transparent mirror 28 is configured such that the laser beam 22 may be directed towards the scanning unit 24 through the mirror 28, and such that thermal radiation directed in the opposite direction is reflected to the sensor 6. Thus, the laser beam 22 and the heat radiation share the same optical path (between the raw material powder and the half mirror 28).

However, it is not necessary to arrange the sensor 6 in the same irradiation device 18 as the laser beam source 20. The sensor 6 may be provided as a separate device with a separate beam path and a separate optical component, e.g. a scanning unit. In the present disclosure, the thermal radiation emitted by the melt pool is considered as a process condition within the apparatus 2. In addition or as an alternative to the sensor 6 for detecting thermal radiation, different sensors for detecting different process conditions may be provided within the detection device 2, such as a temperature sensor for detecting the temperature within the manufacturing chamber 10, a laser power sensor for detecting the laser power of the laser beam 22, and/or an inert gas pressure sensor for detecting the inert gas pressure, etc.

The sensor 6 detects the thermal radiation and outputs a sensor signal. According to this embodiment, the sensor signal is sampled at a known predetermined sampling rate and converted to an analog signal that is a digital signal. The sampling may be performed by the sensor 6 itself or by the control unit 8. The sensor data obtained in either case comprises a time series of data values that can be processed by the control unit 8. According to the present embodiment, the respective data values represent the intensity of the heat radiation detected by the sensor 6. However, the sensor 6 or the control unit 8 may also receive the intensity signal generated by the sensor 6 and generate temperature data comprising a data value indicative of the melt pool temperature based on the intensity signal.

The respective data values may represent the absolute values of the calibrated sensor 6. However, since the data values are compared with each other, it is not necessary to use the calibration values.

Fig. 2 shows an apparatus 30 for analyzing sensor data of a sensor provided in an apparatus for manufacturing a three-dimensional workpiece by irradiating a raw material layer with an energy beam according to the present disclosure. The control unit 8 shown in fig. 1 is an embodiment of the device 30 shown in fig. 2. Thus, the apparatus for manufacturing a three-dimensional workpiece is the apparatus 2 for manufacturing the three-dimensional workpiece 4 shown in fig. 1, and the energy beam is the laser beam 22.

The device 30 of fig. 2 comprises a plurality of units 32 to 42, each of which may be implemented in hardware or software or a combination thereof. For example, the device 30 may comprise a processor and a memory, wherein the memory comprises instructions representing the respective units 32 to 42.

The apparatus 30 comprises a first receiving unit 32, the first receiving unit 32 being configured to receive the sensor data as a time sequence of data values, wherein each data value is indicative of a process condition within the device during manufacturing of the three-dimensional workpiece. The apparatus 30 comprises a second receiving unit 34, the second receiving unit 34 being configured to receive planning data for the three-dimensional workpiece, the planning data defining a plurality of scan vectors and an order in which the energy beam is scanned along the plurality of scan vectors. The device 30 comprises an associating unit 36, the associating unit 36 being configured to associate the data values of the time series with respective vector data of the scan vectors of the planning data to form a plurality of sets of data values of the respective scan vectors. The device 30 comprises a defining unit 38, the defining unit 38 being configured to define a group of at least two scan vectors of the plurality of scan vectors based on the planning data, wherein the scan vectors of the group of scan vectors fulfill a predetermined similarity criterion. The device 30 comprises a comparison unit 40, the comparison unit 40 being configured to compare the set of data values of a first one of the set of scanning vectors with the set of data values of at least one second one of the set of scanning vectors, or to compare the set of data values of the first one of the set of scanning vectors with the combined set of data values deduced from the at least two second one of the set of scanning vectors. The apparatus 30 comprises a determination unit 42, the determination unit 42 being configured to determine a quality metric of the workpiece at the position of the first scan vector based on the comparison.

The above-described units of the device 30 are configured to perform a method comprising:

-receiving sensor data as a time series of data values, wherein each data value represents a process condition within the apparatus during the manufacturing of the three-dimensional workpiece;

-receiving planning data for a three-dimensional workpiece, the planning data defining a plurality of scan vectors and an order in which the energy beam is scanned along the scan vectors;

-associating the data values of the time series with respective vector data of a scan vector of the planning data to form a plurality of sets of data values of the respective scan vector;

-defining a group of at least two scan vectors of the plurality of scan vectors based on the planning data, wherein the scan vectors of the group of scan vectors fulfill a predetermined similarity criterion;

-comparing the set of data values of the first one of the set of scanning vectors with the set of data values of at least one second one of the set of scanning vectors, or comparing the set of data values of the first one of the set of scanning vectors with the combined set of data values deduced from at least two second one of the set of scanning vectors; and

-determining a quality metric of the workpiece at the location of the first scan vector based on the comparison.

In the following, the individual units of the device 30 shown in fig. 2 will be described in detail in the context of the embodiment of fig. 1, wherein the control unit 8 represents the device 30.

A first receiving unit 32 of the device 30 receives the sensor data as a time sequence of data values, wherein the data values represent the intensity of the thermal radiation emitted by the melt bath. Since the sensor 6 is sampled at a known sampling frequency, time information may be associated with each data value. In other words, when, for example, the time at which the first data value was recorded is known, the times of the other data values may be calculated. Accordingly, a respective timestamp may be assigned to one or more of the plurality of data values. It is also possible to deduce certain events from the data values and thus obtain information about the time at which these values were recorded (e.g. turn the laser on, turn the laser off, etc.).

The second receiving unit 34 of the device 30 receives the planning data. The planning data may be stored in a memory of the control unit 8 and may be used by the control unit 8 to obtain control data of the various components of the apparatus 2. In other words, the planning data comprise all information about one particular manufacturing process (i.e. the manufacturing process of the respective workpiece to be manufactured) so that the device 2 can irradiate the desired locations of the individual layers of raw material powder. To this end, the planning data includes data defining a plurality of scan vectors and an order in which the laser beam 22 is scanned along the scan vectors. In the example described below with reference to fig. 3, the scanning vector is a plurality of straight lines having a predetermined length and arranged in parallel with each other to fill the inside of the workpiece 4. Furthermore, a scan vector is provided which represents the profile of the workpiece 4.

Since the measured times of the sensor data are known (see above), the individual data values of the time series may be associated (by the association unit 36) with the respective vector data of the scan vectors of the planning data. As a result, a plurality of sets of data values are generated, wherein one set of data values corresponds to one of the plurality of scan vectors. These sets of data values are stored in association with information linking the respective sets of data values with the respective scan vectors.

The defining unit 38 defines a group including at least two scan vectors satisfying a predetermined similarity criterion among the plurality of scan vectors. The operation is performed based on the planning data without taking into account the corresponding sensor data. Thus, the operation may also be performed at an earlier point in time, i.e. before the sensor data is received. In general, the order of the method steps is not limited to the above-described order, and the method may be performed in any order as long as it is technically possible.

The similarity criterion may be any criterion suitable for defining the similarity between scan vectors. For example, scan vectors having a particular length (or having a length within a particular range of lengths) may be grouped by a qualifying unit, while scan vectors having different lengths (or having a length within a particular, different range of lengths) may be grouped into different groups. According to different examples, all scan vectors with a particular direction may be grouped together. Thus, all scan vectors pointing to the right (i.e., having a component of positive x or zero x with respect to an arbitrarily selected x-axis) are grouped into a first group, while the remaining scan vectors pointing to the left (i.e., having a negative x-component) are grouped into a second group. Other criteria for defining similarity may be based on, for example, the position of the scan vector relative to the workpiece 4 (e.g., the scan vector is part of the contour of the workpiece, part of the interior region of the workpiece, part of the up/down/left/right side regions of the workpiece).

The comparison unit 40 compares the set of data values of the first scan vector of the set of scan vectors with the set of data values of the at least one second scan vector of the set. Alternatively, the set of data values of the first scanning vector is compared with a combined set of data values derived from at least two second scanning vectors of the set. While this disclosure describes the comparison of scan vectors to one another, it should be understood that the comparison step described above means comparing the respective sets of data values of the respective vectors to one another.

According to one embodiment, each newly measured scan vector is compared to previously recorded scan vectors in the same group. As mentioned above, this means that the respective sets of data values of the scanning vectors are compared with each other. For example, the respective types of distance metrics may be calculated for the two vectors. The distance metric may be any distance metric suitable for identifying a data value that may be indicative of the varying quality of the workpiece at the location of the corresponding scan vector.

An example of the operation of comparing and determining the distance metric is described below.

-determining a difference between a maximum of the set of data values of the first vector and a maximum of the set of data values of the second vector;

-determining a difference between a minimum of the set of data values of the first vector and a minimum of the set of data values of the second vector;

-determining a difference between the average of the set of data values of the first vector and the average of the set of data values of the second vector;

-determining a difference between a median of the set of data values of the first vector and a median of the set of data values of the second vector; or alternatively

-determining the sum of the absolute differences between the respective data values of the two sets.

The comparison unit 40 then determines whether the distance metric exceeds a threshold. The threshold may be a predetermined (fixed) threshold. The threshold may also be derived based on previously measured distance metrics in the same group. For example, the threshold is the average of previously measured distance measures or the average multiplied by a predetermined factor. According to a different example, the threshold value corresponds to a distance metric measured directly before the current distance metric, or to the distance metric multiplied by a predetermined factor (e.g., 1.1, 1.2, 1.5, or 2.0). According to this example, a current scan vector of a set of scan vectors is compared to a previous scan vector of the set, and the comparison result is compared to a previous comparison result, wherein the previous comparison result is a comparison result between the previous scan vector and a scan vector preceding the previous scan vector. In this way, it may be determined whether the set of data values of the current scan vector is within the normal range of the corresponding set. In the case where the data values are significantly different, it can be judged that a quality problem has occurred at the position of the corresponding vector.

As an alternative to comparing each newly measured scan vector with previously recorded scan vectors in the same group, a combined set of data values derived from at least two second scan vectors of the group may be generated. The data value set of the newly measured scan vector (first scan vector) is then compared to the combined data value set. In other words, a distance measure between the combined set of data values and the set of data values of the first scanning vector is calculated. In case the distance measure is larger than a predetermined threshold, the determination unit 42 determines that the quality measure of the workpiece at the position of the first scan vector is lower than the quality measure of the workpiece at other positions of the workpiece. According to one embodiment, the combined set of data values may be an averaged set of data values. More precisely, an averaging form may be applied to the set of data values of the at least two second scanning vectors in the group to derive an averaged set of data values. In other words, the set of combined data values may represent an average vector, i.e. an average version of the data values of the combined vector. Averaging may comprise deriving an arithmetic mean of the data values of the different vectors. The combined set of data values may be deduced by considering at least one of meridians, extrema and quantiles.

The operation of the comparison unit 40 may also be performed as follows. In accordance with one or more embodiments, comparisons between absolute intensity processes (drop/increase in intensity from vector to vector) as well as relative comparisons (e.g., differences/deviations/fluctuations of vectors from each other) are possible. Together with the planning data, conclusions can be drawn about the quality of the workpiece (e.g. the position of the binding vector within the layer).

The comparison result may be determined according to any suitable quality criterion or numerical measure, such as correlation, absolute difference, extremum analysis, etc. By additionally taking into account information of system parameters or planning data, a quality criterion of the multidimensional structuring can be formulated. These associations may be related to each other through a neural network or a decision tree.

All quality data can be collected and stored for later use. The resulting (increasing) database can be used for "learning", i.e. for which scan vector which workpiece quality is desired. These values may be used for reference values (or more precisely, sets of reference data values) of scan vectors of different classes (i.e. different groups) for subsequent scans. As an example, an algorithm based on (e.g., random) sample value "learning" may be implemented to reduce the amount of data processed and stored. For reference, each previous manufacturing process may be used.

The determination unit 42 determines a quality measure of the workpiece 4 at the position of the first scan vector based on the comparison result of the comparison unit 40. In case the comparison result shows that the set of data values of the first scanning vector differs substantially from the compared set of data values (i.e. exceeds a threshold value), the determination unit determines that the quality measure of the workpiece at the position of the first scanning vector is degraded.

The location of the "problematic" first scan vector may be stored and output to the user. In this way, a user may obtain information regarding potential problem areas of the workpiece (i.e., areas where the quality of the workpiece may have changed).

Additionally or alternatively, the control unit 8 of the device 2 may initiate countermeasures to improve the quality of the irradiated scan vectors. In other words, process parameters of the device may be adjusted during the manufacturing process based on the determined quality metric. For example, irradiation parameters such as focus position or laser power may be adjusted.

The operation of the determination unit 42 may also be performed as follows. One example is the effect of a deviation in the development of the intensity signal. These deviations can be correlated with the waviness of the respective outer contour, which can be understood as a classifier of the surface quality. In the case of excessively high laser powers, bubbles and evaporation can occur, which again leads to a strong waviness of the outer contour and thus to a corresponding reduction in the quality of the workpiece.

Furthermore, in a multi-optical system, the time course between different optical components and the influence on the workpiece quality can be understood. For example, a significantly increased intensity profile is obtained when two lasers irradiate adjacent areas simultaneously or at similar times.

The similarity measure of the vector comparison may be later attributed to a density measurement of the finished workpiece, or a crystalline structure, tensile strength, or different quality characteristic of the workpiece.

Also, multidimensional methods are contemplated. In each layer, the respective group (or class) intensities may be displayed as a 2D intensity image. These 2D images may be arranged for different irradiation layers and may be viewed and evaluated as a 3D data set. These 3D data sets may be analyzed as N-dimensional data sets in different manufacturing processes.

In this way, chronological information of the layers and the correlation (e.g. information about the height of the workpiece or the amount of powder in the manufacturing cylinder) can be obtained.

Fig. 3 shows an example of a layer forming a decagonal workpiece 4 filled with a plurality of scan vectors 44. In fig. 3, for greater clarity, only two exemplary scan vectors are indicated with reference numerals. More specifically, the interior of the layer of the workpiece is filled with a plurality of scan vectors 44 defining a pattern according to which a majority of the plurality of scan vectors 44 have a predetermined fixed length and the scan direction of the scan vectors is reversed after each scan vector.

In the example shown in FIG. 3, multiple columns of scan vectors are scanned, where the columns of scan vectors extend along the y-axis and each scan vector is parallel to the x-axis. In one exemplary column (the middle column shown in fig. 3), a first scan vector V1 is irradiated along the positive x-direction, followed by a second scan vector V2 being irradiated along the negative x-direction. Thereafter, a third scan vector V3 is scanned along the positive x-direction, and so on. In this way, the columns are scanned along the positive y-direction as indicated by the thick arrows in fig. 3.

The scanning strategy described above is defined by the planning data of the three-dimensional workpiece 4 to be manufactured. The planning data is analyzed to define different sets of scan vectors that satisfy certain similarity criteria. As an example shown in fig. 3, scan vectors extending along the positive x-direction are associated as a first group and scan vectors extending along the negative x-direction are associated as a second group.

Furthermore, as described above, during irradiation of the scan vector 44, the sensor 6 measures the intensity of thermal radiation emitted by the melt pool. The generated data values are associated with corresponding vector data of the scan vector.

As shown in fig. 3, a pair-wise comparison is made between respective sets of data values of a plurality of scan vectors. When the third scan vector V3 has been scanned, the corresponding set of data values of the third scan vector is compared with the set of data values assigned to the first scan vector V1. Similarly, V4 is compared to V2, V5 is compared to V3, V6 is compared to V4, and so on.

Thus, the set of data values of a particular scan vector is always compared to the set of data values of the scan vectors of the same group. The data values of similar scan vectors may be compared to each other. Thus, in case there is no quality problem for the respective scan vectors, the result of the comparison should show that the data values are similar.

FIG. 4 shows different scan vectors V that can be compared with each other₁To V₃An example of a data value of (c). Since the individual data values have already been assigned a time stamp, the data values can be plotted on a (horizontal) time axis. The vertical axis represents the corresponding data value, i.e. the intensity value of the intensity of the thermal radiation emitted by the melting bath(arbitrary unit).

Based on the representation of fig. 4, it is clear that the data values of the vectors can easily be compared to each other (e.g. by comparing extreme values) according to one of the above-described techniques.

Furthermore, the double-headed arrows in fig. 4 indicate that only a subset of the set of data values (and not all data values) of each scan vector may be compared to each other, according to one embodiment. In this case, subsets of fixed duration are defined, and only the data values of the respective portions of the subsets of each scan vector are compared with each other. In this way, a region of the scan vector in which there are no data values for human error can be selected. Furthermore, in case the mutually compared scan vectors have different lengths, it is necessary to select subsets such that subsets having the same length are compared with each other (or a subset of one of the vectors is compared with the entire other vector having the same length as the subset).

It should also be generally noted that the two or more scan vectors (i.e., their respective sets of data values) that are compared to each other are not necessarily scan vectors that extend in the same direction. The similarity criteria may also group vectors that generally extend in opposite directions or different directions. In the example of fig. 4, two scan vectors V may be used₁And V₂(or respective subsets thereof) are compared to each other even though the two scan vectors may have been scanned in opposite directions (in the case of a meandering scanning strategy).

Fig. 5 shows that the scan vectors of the different layers (L1 to L4) can also be compared with each other. This fig. 5 shows a plurality of layers arranged on top of each other, wherein the cross-section of each layer is again, by way of example, a decagon as shown in fig. 3. According to the embodiment shown in fig. 5, a plurality of scan vectors that are part of the workpiece profile are set as a group. Furthermore, the measured data values of the scan vectors are compared with the data values of the scan vectors of different (e.g., adjacent) layers. As shown in fig. 5, the scan vector of the profile of L1 is compared with the scan vector of the profile of L2. Similarly, the scan vector of the profile of L2 is compared to the scan vector of the profile of L3, and so on. Comparison of scan vectors between different layers may also be performed for scan vectors that are not part of the workpiece profile (i.e., vectors for interior portions of the workpiece). Further, the scan vector may be compared to a scan vector of a previous manufacturing process (e.g., for a series production of the same manufacturing instruction).

In the above description, embodiments have been described that provide an improved technique of analyzing sensor data of a sensor provided in an apparatus for manufacturing a three-dimensional workpiece by irradiating a raw material layer with an energy beam.

Claims

1. An apparatus for analyzing sensor data of a sensor (6), the sensor (6) being arranged in a device (2) for manufacturing a three-dimensional workpiece (4) by irradiating a layer of a starting material with an energy beam (22), the apparatus comprising a control unit (8), the control unit (8) being configured to:

receiving the sensor data as a time series of data values, wherein each data value represents a process condition within the apparatus (2) during the manufacturing of the three-dimensional workpiece (4);

receiving planning data for the three-dimensional workpiece (4), the planning data defining a plurality of scan vectors (44) and an order in which the energy beam (22) is scanned along the scan vectors (44);

associating the time series of data values with respective vector data of the scan vectors (44) of the planning data to form a plurality of sets of data values for respective scan vectors (44);

defining a group of at least two scan vectors (44) of the plurality of scan vectors (44) based on the planning data, wherein the scan vectors (44) of the group satisfy a predetermined similarity criterion;

comparing the set of data values of the first scanning vector (44) in the set with the set of data values of at least one second scanning vector (44) in the set, or comparing the set of data values of the first scanning vector (44) in the set with a combined set of data values derived from at least two second scanning vectors (44) of the set; and

determining a quality metric of the workpiece (4) at the location of the first scan vector (44) based on the comparison.

2. The apparatus of claim 1, wherein the data values are intensity values or temperature values of thermal radiation generated in a melt pool in which the energy beam (22) impinges on the raw material.

3. The apparatus of claim 1 or 2, wherein the apparatus is configured to perform at least the steps of receiving the sensor data, correlating, comparing and determining during a manufacturing process of the workpiece (4).

4. The apparatus of claim 3, wherein the control unit (8) is further configured to adjust a process parameter of the device (2) during the manufacturing process based on the determined quality metric.

5. The device according to any one of claims 1 to 4, wherein the control unit (8) is further configured to associate time data with the time series of data values.

6. The apparatus of any of claims 1-5, wherein the similarity criterion takes into account at least one of: -a length similarity of the scan vectors (44), -a direction similarity of the scan vectors (44), and-whether the scan vectors (44) are part of the contour of the workpiece (4).

7. The device according to any one of claims 1 to 6, the control unit (8) being further configured to compare a set of data values of the first scanning vector (44) in the group with a predetermined set of reference data values associated with the group.

8. The device according to any one of claims 1 to 6, wherein the control unit (8) is configured to:

comparing the set of data values of the first scan vector (44) in the set with the set of data values of a previously irradiated second scan vector (44) in the set; and

determining at least one difference value based on the comparison, the at least one difference value representing a difference between the set of data values of the first scan vector (44) and the set of data values of the second scan vector (44).

9. The apparatus according to claim 8, wherein the control unit (8) is further configured to:

comparing the at least one difference value to a stored difference value representing a difference between sets of data values of two previously irradiated scan vectors (44).

10. The device according to any one of claims 1 to 9, wherein the control unit (8) is configured to perform the step of comparing by taking into account at least one of:

a series of absolute data values, e.g. a decrease and/or an increase of data values from scan vector (44) to scan vector (44);

relative comparisons of data values, such as differences, deviations and/or fluctuations between the data values;

-a correlation between a set of data values of the first scanning vector (44) and a set of data values of the at least one second scanning vector (44), or a correlation between a set of data values of the first scanning vector and a combined set of data values deduced from the at least two second scanning vectors (44);

an absolute difference between a data value of the first scanning vector (44) and a set of data values of the at least one second scanning vector (44), or an absolute difference between a data value of the first scanning vector and a combined set of data values derived from the at least two second scanning vectors (44); and

an extreme analysis of an extreme of the set of data values of the first scanning vector (44) and an extreme of the set of data values of the at least one second scanning vector (44), or an extreme analysis of an extreme of the set of data values of the first scanning vector and an extreme of the combined set of data values deduced from the at least two second scanning vectors (44).

11. The apparatus according to any one of claims 1 to 10, wherein the control unit (8) is configured to: the step of comparing is performed by considering only a subset of the set of data values of the first scanning vector (44) and/or a subset of the set of data values of the at least one second scanning vector (44), or considering a subset of the combined set of data values deduced from the at least two second scanning vectors (44).

12. The apparatus of any one of claims 1 to 11, wherein the control unit (8) is configured to output a data set representing at least one two-dimensional image of a layer of the workpiece (4), wherein the time-series of data values are assigned to corresponding pixels of the two-dimensional image.

13. An apparatus (2) for manufacturing a three-dimensional workpiece by irradiating a layer of raw material with an energy beam, the apparatus (2) comprising:

the apparatus of any one of claims 1 to 12;

an energy beam source (20) for generating the energy beam (22) and irradiating the energy beam (22) onto the layer of raw material; and

a sensor (6) configured to measure and transmit a time series of data values to the device.

14. A method for analyzing sensor data of a sensor (6) arranged in an apparatus (2) for manufacturing a three-dimensional workpiece (4) by irradiating a layer of a starting material with an energy beam (22), the method comprising:

15. A computer program product stored on a computer readable storage medium, comprising computer readable instructions for causing a computer to perform the method of claim 14.